NO309309B1 - Device for separating leukocytes from platelet concentrate - Google Patents

Description

The present invention relates to a device for reduction

of the leukocyte content of platelet concentrate derived from blood.

Development of plastic blood collection bags in the 1960s facilitated the separation of donated whole blood into its various components, thus making platelet concentrates available as a transfusion product. Separation of a single unit of donated whole blood, approximately 450 ml according to US practice, into its individual components is accomplished using differential sedimentation. The unit of whole blood is collected in a plastic blood collection bag with integral satellite transfer bags, and is separated by centrifugation into a concentrate of red cells and platelet-rich plasma. The platelet-rich plasma is transferred to an empty associated satellite bag, and the sedimented red cells are separated from the collection system. The platelet-rich plasma is centrifuged at an increasing G force which separates it into platelet concentrate (PC) and platelet-poor plasma, which are then separated by transfer to a satellite bag. The final PC should contain, on average, no less than 5.5 x 10<10> platelets in 50 to 70 ml of plasma, or about 10^ platelets per microliters (ul). Separation of the components must take place within 6 hours of collection of the whole blood. One unit of platelet concentrate usually increases the platelet count of a 70 g individual by 5000-10,000/ul. The usual dose administered to an adult with thrombocytopenia varies from 6-10 units of concentrate.

Platelets can also be prepared using a specialized blood component collection procedure termed "apheresis". Using a variation between continuous or discontinuous flow devices, the platelets are collected from a single donor. Using this method, whole blood is removed from the donor, and then it is centrifuged using the ex-vivo technique into the individual components. The desired component, in this case platelets, is harvested, and the remainder of the autologous blood is returned to the donor. This method allows collection of multiple units from one donor. Typically, the 2 to 3 hour apheresis procedure will produce a platelet product containing 3 x IU platelets, which is equivalent to 6-8 units of platelets from random donors. Donors who provide single-donor platelet concentrates are usually HLA-matched with the recipient. Platelets from single donors are usually given to patients who are immunologically unresponsive to transfusions of platelet concentrates from random donors, or to individuals who are considered candidates for bone marrow transplantation.

Platelets fulfill two main functions in maintaining hemostasis. First, they adhere to damaged blood vessel walls, aggregate at the damaged site, and form a hemostasis plug. Secondly, they participate in the formation of fibrin by releasing platelet factor 3 and by promoting coagulation factor-mediated hemostasis. Platelets also release vasoactive amines, cationic proteins, nucleotides and enzymes, as well as thromboxane A2 which induces vasoconstriction and promotes platelet aggregation via its inhibitory effect on the formation of cyclic AMP.

Platelet transfusions are for the treatment of bleeding caused by thrombocytopenia secondary to insufficient production of platelets in the bone marrow, a condition known as amegakaryocytic thrombocytopenia. This bone marrow hypoplasia can be caused by chemotherapy, tumor invasion, or primary aplasia. For example, a patient with acute leukemia may be thrombocytopenic at diagnosis or become thrombocytopenic secondary to chemotherapy or radiation therapy procedures. Patients with adequate platelet counts but with abnormal congenital platelet function may have a variety of reactions resulting in mild to severe hemorrhagic disease, such as Glanzmann's thrombosthenia gravis. Patients may also exhibit functional platelet disorders secondary to plasma abnormalities, such as von Willbrand's disease or uremia. Platelet transfusions can also be used in patients suffering from thrombocytopenia associated with massive blood replacement secondary to the trauma, or in patients exposed to surgical interventions that require large amounts of blood, e.g. open heart surgery. Patients who have ingested aspirin may also demonstrate transient platelet dysfunction and require platelet transfusion support in emergency surgical procedures.

The use of platelet concentrates has increased and is still increasing rapidly. This is demonstrated by the latest survey conducted by "The American Blood Commission". This survey shows that the national use of platelet concentrates increased from 0.41 million in 1971 to 2.86 million in 1980, which corresponds to a 6-fold increase. In contrast, the use of packed red cells (PRC) during the same period was from 6.32 million to 9.99 million units, which corresponds to only a 1.5-fold increase. Platelets from random donors were prepared from less than 6 percent of the whole blood units collected in 1971, compared with almost 20 percent in 1980. The proportion of donated whole blood units used to prepare the platelet concentrate in 1987 is approximately 70-80 percent. The needs of the future may exceed the available blood supply, and the needs that exist today already exceed the available blood supply in some places.

There are several reasons for this increased use of platelets, including a more aggressive use of chemotherapy with resulting prolonged periods of bone marrow aplasia. The availability of platelet components and the more aggressive utilization of platelet transfusions allow the use of these aggressive chemotherapy regimens.

Transfusion of the platelet concentrate is not without risk for those patients who receive both acute and chronic transfusion support. Chills, fever and allergic reactions can occur in patients receiving acute as well as chronic platelet therapy. Repeated platelet transfusions often lead to alloimmunization against HLA antigens, as well as against platelet-specific antigens. This in turn reduces susceptibility to platelet transfusion. The leukocytes that contaminate the platelet concentrates, including granulocytes and lymphocytes, are associated with both febrile reactions and alloimmunization leading to platelet transfusion refractoriness. Another life-threatening phenomenon that affects severely immunosuppressed patients is "Graft versus host disease". In this clinical syndrome, donor lymphocytes transfused with platelet preparations can initiate an immunological reaction against the host, i.e. the transfusion recipient, with pathological consequences. Another potential consequence of platelet transfusion is the transmission of bacterial, viral and parasitic infectious diseases.

There is much evidence that leukocyte-coated platelet concentrates reduce the incidence of febrile reactions and platelet refractoriness. Leukocyte-drained blood components should also be evaluated for a potential role in reducing the potential for Graft vs. the cough disease. Leukocyte depletion of platelet preparations may also reduce the transmission of certain viruses (eg Acquired Immunodeficiency Syndrome (AIDS) and cytomegalovirus (CMV)).

The platelet preparations contain varying amounts of leukocytes. The platelet concentrates prepared by differential centrifugation of the blood components have varying leukocyte contamination which is linked to the centrifugation time and the G strength used. Leukocyte contamination is also influenced by the choice of apheresis technique used to harvest the component. While the dose of the contaminating leukocytes required to cause a febrile reaction or trigger platelet refractoriness with repeated transfusions remains unknown, Stec et al (Stec, N. , Kickler, T.S. , Ness, P.M. and Braine, H.G., Effectiveness of Leukocyte ( WBC) Depleted Platelets in Preventing Febrile Reactions in Multi-Transfused Oncology Patients, American Association of Blood Banks, San Francisco, (Abstract No. 598), November 3-7, 1986), and Dan and Stewart (Dan, M.E., and Stewart , S., Prevention of Recurrent Febrile Transfusion Reactions Using Leukocyte Poor Platelet Concentrates Prepared by the "Leuko-trap" Centrifugation Method, American Association of Blood Banks, San Francisco, (Abstract No. 597), November 3-7, 1986 ) demonstrated that leukocyte removal efficiencies of 81 to 85$ are sufficient to reduce the incidence of febrile reactions to platelet transfusions. Several other studies have recently reported a reduction in alloimmunization and platelet refractoriness at leukocyte contamination levels of <1 x IO<7> per unit. The leukocyte contamination level in conventional platelet preparations is generally >1 x IO<8. >The existing studies therefore suggest that at least a two-log (99%) reduction of leukocyte contamination is required. Later studies have suggested that a three log ($99.9) or four log ($99.99) reduction would be considerably more beneficial. A further desired criterion is to limit platelet loss to approximately 50$ or less of the native platelet concentrate.

Centrifugal methods are available that reduce the number of leukocytes contaminating the platelet preparations. These methods have often been shown to be unsatisfactory because they result in unacceptable platelet loss. Centrifugation, along with the use of specially designed pooling bags, reduces the concentration of leukocytes by about a single log. However, this technique is expensive and labor-intensive.

The use of laboratory filters to remove contaminating leukocytes from platelet preparations has, in some cases, provided removal efficiencies of 2 log leukocyte with a platelet yield 1 average of 90$; however, in most studies using laboratory filters, unacceptably high platelet losses have been obtained. Experience with the use of laboratory filters to clear platelet preparations of leukocytes has shown that the methods are not accurate when performing these. Furthermore, use of these devices is labor-intensive and results in reduced storage time for a unit of conventionally collected platelets. Due to the reduced storage time, units prepared using such devices are not recommended for bedside use.

Characteristics that are desirable in a leukocyte collection device;

The ideal device for draining leukocytes from platelet preparations should be readily available, inexpensive, designed for bedside use, and have the ability to deliver the platelet components to the patient within 30 seconds of initiation of transfusion of the platelet preparation. The leukocyte content of the platelet preparation should be reduced by means of the device by at least 99$ or to a total amount not greater than 5 x 10^ leukocytes per administration and preferably with 99.9$ or more to a level of less than 5 x 10^ leukocytes per administration. After infusion into a patient, platelet function should be only minimally affected, and platelet survival time within the new host should be near normal. Because of the high expense and increased demand for platelet preparations, as well as the clinical necessity that includes delivery of a maximum therapeutic dose, this ideal device should deliver the highest possible amount of the platelets initially present in the bag. Such a device is an object of the present invention.

Devices which have previously been used in attempts to fulfill this purpose have been based on the use of packed fibers and have generally been referred to as "filters". When describing this invention, the terms "tapping device", "element", "composition", "filter" and "filter-adsorbent" are used interchangeably.

Extraction of platelets:

Platelets are notorious for being "sticky" which is a term that reflects the tendency of platelets suspended in blood plasma to adhere to any non-physiological surface to which they are exposed. In many cases, they also strongly adhere to each other.

Platelets are also sensitive to various environmental stimuli, and one includes exposure to cold. While the blood banks store other blood components at 6°C or less to extend the useful life, platelets are best preserved at normal indoor room temperature, e.g. 20° to 22°C. At this temperature, their nominal useful life in the U.S. practice 5 days, although many medics prefer to apply them during the first 2 or 3 days after collection.

While the number of red cell concentrate units normally infused into a patient is one per administration, the usual practice with regard to platelet concentrate is to transfuse a mixture of six to ten units of platelets per administration, containing a total of approximately 300 to 700 ml of platelet concentrate.

Based on 1987 U.S. prices in the U.S. dollars, includes a 1% improvement in platelet recovery for a savings of more than about four dollars. Regardless of the sum of money, high recovery efficiency is important because platelet concentrate is often in short supply.

An object of this invention is a leukocyte tapping device with the highest possible platelet recovery.

Recovery of the platelets can be adversely affected in several ways: (a) The platelets can adhere to the surfaces of components of the leukocyte collection device. If the device depends on filtration or adsorption as a mechanism for leukocyte removal, the internal surface area of the filter (eg, the surface area of the fibers in a fibrous filter, or the fiber surface area) will be substantially on the order of one or more square meters, and adhesion of the platelets to the fiber surfaces tends to cause substantial, and often almost complete, removal of the platelets. (b) Platelet concentrate present within the leukocyte draining device upon completion of transfusion will be lost. For this reason, the internal volume of the device should be as small as is compatible with providing the desired degree of leukocyte depletion. (c) The internal volume of the extra parts of the appliance,

including the piping and drip tray should be as small as possible.

The leukocyte collection device according to this invention directly and effectively minimizes the platelet loss caused by the above-mentioned causes.

The viability of the platelets after leukocyte collection:

In any system that depends on filtration to remove leukocytes, there will be substantial contact between the platelets and the inner surfaces of the filter. The filter must be such that the platelets are not adversely affected by this contact. Such a filter is an object of this invention.

Capacity:

As with separation from whole blood in current blood banking practice, the platelet concentrate not only contains a large amount of leukocytes present in the native blood, but it may also contain fibrinogen, fibrin threads, small fat globules, some red cells and other components usually present in small amounts . During the centrifugation process that concentrates the platelets and partially separates them from the remaining components, there is a tendency for microaggregates to form. These microaggregates may include some platelets along with leukocytes, red cells, fibrin and other components. Gels that can be formed from fibrinogen and/or fibrin are present in significant amounts in platelet concentrates prepared using U.S. Pat. permitted blood banking methods.

The platelet concentrate (PC) is stored with continuous gentle mixing at approximately 20-22°C for use over a period of 5 days. Mixing prevents agglomeration, encourages gas exchange (thus controlling pH), and bathes the product in necessary nutrients. Despite this, the number and size of microaggregates increases with time. Furthermore, gel-like bodies may be formed, which may include fibrinogen, degenerate protein and degenerate nucleic acids.

If the leukocyte collection device includes a porous structure, the microaggregates and gels tend to accumulate on or within the pores, causing blockages that in turn inhibit flow. Normally, gravity is used in transfusions, which develop more than about 0.1 to 0.14 kg/cm 2 to induce the flow from the storage bag through the leukocyte collection device to the patient. Because of this, a particularly important characteristic of the separation device includes that it is resistant to clogging.

Because of the unusual and highly variable combinations of sealing factors, the experience of a person trained in the field of filter construction is insufficient when applied to remove unwanted components from the platelet concentrate, and new, inventive approaches have been required to construct a filter that on an efficient manner retains leukocytes, allows a high percentage of platelets to be passed through, and will reliably pass up to 10 units of platelets without clogging. Development of such a device is a purpose of this invention.

The lightness and speed at start:

Ease of use is an important characteristic of any leukocyte collection system. As indicated above for leukocyte tapping devices, an easy start initiation is a particularly important factor. The term "start" refers to the initiation of the flow of the platelet concentrate from the bag through the filter and into the patient. A short start-up period is always desirable to save nursing/technician time. An object of this invention is to keep this time to less than about 10 to 20 seconds when starting by gravity.

Pretreatment of the leukocyte collection device prior to initiation: A number of currently used devices require pretreatment before the PC is passed through, this usually involves the passage of physiological saline, some of which may be delivered into the patient's vein. The necessity for such an operation is clearly very undesirable for the reasons stated above. The reasons for using such pretreatments vary, and these reasons include removal of acid hydrolyzate developed during the steam sterilization of the devices containing cellulose acetate fibers, assurance of the absence of extraneous solids that may be present in natural fibers, and prevention of hemolysis (loss of the integrity of the red blood cells with subsequent loss of their contents to the external environment) if the fibers are hygroscopic. When synthetic fibers are used, sending saline as a first step helps the problems caused by poor wettability of the synthetic fibers, which can result in parts of the fibrous medium remaining unmoistened during the leukocyte draining process. If significant portions of the medium remain unmoistened, the pressure drop is higher and less of the fiber surface is available for leukocyte removal, thus causing a lower efficiency to be provided.

The invention provides a leukocyte collection device that requires no pretreatment prior to bedside application.

Wetting the fibrous medium:

When a liquid is brought into contact with the surface upstream of a porous medium and a small pressure differential is used, flow into and through the porous medium may or may not occur. A condition in which no flow occurs is in which the liquid does not wet the material of which the porous structure is made.

A series of liquids can be prepared, each with a surface tension about 3 dynes/cm higher than the previous one. A drop of each can then be placed on a porous surface and one can then observe to determine if it is absorbed quickly or if it remains on the surface. For example, solid PTFE has a critical surface tension ("Yc) of 18 dyne/cm, and thus, by definition, is not wetted by a liquid with a surface tension greater than 18 dyne/cm. In contrast, wetting is observed when apply the droplet technique to a 0.2 µm porous polytetrafluoroethylene (PTFE) filter sheet for a liquid droplet with a surface tension of 26 dyne/cm, but the porous surface remains unwetted when a liquid droplet with a surface tension of 29 dyne/cm is applied. the lower surface tension is spontaneous upon contact - no pressure, vacuum or other manipulation is required.

Similar behavior is observed for porous media produced using other synthetic resins, in which the wetted-unwetted values depend primarily on the surface characteristics of the material from which the porous medium is produced, and secondarily, on the pore size characteristics of the porous medium . For example, fibrous polyester (specifically polybutylene terephthalate, herein referred to as "PBT") sheets having pore diameters less than about 20 µm will be wetted by a liquid with a surface tension of 50 dynes/cm, but will not be wetted by a liquid with a surface tension of 54 dyne/cm. This contrasts with the critical surface tension (CST) of solid PBT, which is approximately 44 dyne/cm.

To characterize this behavior of the porous medium, the term "critical wetting surface tension" (CWST) has been defined as described below. The CWST of a porous medium can be determined by individually applying a series of liquids with surface tensions varying by 2 to 4 dynes/cm to the surface, and then observing the absorption or non-absorption of each liquid. The CV/ST of a porous medium, in units of dynes/cm, is defined as the mean value of the surface tension of the liquid that is absorbed and the value of a liquid of similar surface tension that is not absorbed. Thus in the examples in the two preceding paragraphs, the CWSTs are respectively 27.5 and 52 dyne/cm. If the surface tension interval is an odd number, for example 3, it is a matter of judgment whether the porous medium may be closer to the lower or higher value, and on this basis, as an example, 2.7 or 28 may be entered for

PTFE.

In measuring CWST, a series of standard liquids for testing is prepared with surface tensions varying in a sequential manner by about 2 to about 4 dynes/cm. Ten drops, each 3 to 5 mm in diameter, of each of at least two of the sequential surface tension standard liquids are placed on representative locations of the porous medium and allowed to stand for 10 minutes. After 10 to 11 minutes, the observation is carried out. "Wetting" is defined as absorption into or wetting of the porous medium by at least nine of the ten droplets within 10 minutes. Non-wetting is defined by non-wetting or non-absorption of two or more drops within 10 minutes. The analysis is continued using liquids of successively higher or lower surface tension, until a pair has been identified, one wetting and one non-wetting. which is most closely related in surface tension. The CWST is thus within that range and, the average of the two surface tensions can be used as a single number to specify the CWST. When the two test fluids differ by 3 dynes, a judgment determines how close the sample is to one or the other, and a whole number is thus entered.

Appropriate solutions with varying surface tension can be prepared in many different ways, but those used to develop the product described here included: Because no stable solutions with CWST >115 dyne/cm at room temperature were found, media with CWST > 115 arbitrarily assigned a CWST value of 116 dyne/cm, and one should note that media with this gradation may have higher real CWST values.

Wetting of fibrous media with platelet concentrate:

In PC, the cells are suspended in blood plasma, which has a surface tension of 73 dyne/cm, thus if PC is placed in contact with a porous medium, spontaneous wetting will occur if the porous medium has a CWST of 73 dyne/cm or higher.

The advantages added to the fibers pretreated to CWST values higher than 73 dyne/cm include:

(a) The time for achieving take-off is reduced.

(b) Synthetic fiber medium where the CWST values have been increased by grafting has superior fiber-to-fiber bonding and is, because of this, preferred for use in the production of the preformed elements used in this invention. (c) A portion of the porous medium may remain unwetted,

after the initialization is complete and the streaming of the PC is started. An adverse effect associated with non-wetting is that the non-wetted parts are not available for leukocyte removal, whereby the efficiency of leukocyte removal by adsorption is reduced, the pressure drop is increased with consequent reduction in flow and clogging may occur

arise.

(d) Devices that have been manufactured using unmodified synthetic fibers should preferably be flushed with saline or other physiological fluid before use. This operation is undesirable because it causes loss of PC due to staying within the more complex piping arrangement required, becomes more expensive, and increases operating time and operating complexity.

the tet, and also increases the probability that the sterility may be lost.

According to the purpose of the invention, a device and method for draining the leukocyte content in a platelet concentrate is provided.

The invention provides a device for emptying the leukocyte content of a platelet concentrate comprising a porous, fibrous medium (12) having a CWST of at least about 90 dyne/cm, preferably of at least about 95 dyne/cm. The fibers of the medium can be modified by the presentation of hydroxyl groups when immersed in an aqueous fluid. These fibers can be modified by exposure to an energy source while in contact with a monomer comprising a polymerizable group (eg, an acrylic or methacrylic moiety) and a hydroxyl-containing group (eg, hydroxyethyl). The monomer may be hydroxyethyl methacrylate.

The article according to the invention can also have intermediate fibers which have been modified to present hydroxyl groups together with a smaller number of another anionic group, wherein the second anionic group includes carboxyl. The fibers of the medium can be modified by exposure under polymerization conditions to a monomer containing a polymerizable group (eg, an acrylic or methacrylic moiety) and a carboxyl-containing group, such as methacrylic acid. The fibers may be modified with a mixture of monomers including methacrylic acid and hydroxyethyl methacrylate wherein the acid/acrylate monomer weight ratio in the modified mixture may be between 0.01:1 to 0.5:1, and the acid/acrylate monomer weight ratio may preferably be between 0.05: 1 to 0.35:1. The modification can be carried out using a grafting solution containing 2 to 10% by weight of tertiary butyl alcohol, preferably 4 to 5% by weight of tertiary butyl alcohol. The modification may also be carried out using a grafting solution containing sufficient water soluble alcohol, or alcohol ether to reduce its surface tension to less than about 40 dynes/cm. The alcohol ether that is used can be diethylene glycol monobutyl ether or ethylene glycol monobutyl ether.

The concentration of hydroxylethyl methacrylate in the modifying mixture may be higher than 0.1%, preferably higher than 0.2% by weight. The hydroxyethyl methacrylate concentration in the modifying mixture may be within the range of from 0.2 to 0.7% by weight.

The object according to the invention further provides a device for reducing the leukocyte content in a platelet concentrate or solution, or treating blood or blood component, which comprises a housing 11 with an inlet 13 and an outlet 14 and a fluid flow path between the inlet 13 and the outlet 14, characterized in that the fluid flow path is a porous, fibrous medium 12 with a CWST of at least 90 dyne/cm and a negative zeta potential.

The object of the invention further provides a device for draining the leukocyte content of a platelet concentrate for use with a mixed amount of 6 to 10 units of platelet concentrate, each of 50 to 70 ml in volume, and in which the effective flow area is greater than 40 square centimeters. The effective flow area of this device can further be greater than 50 square centimeters, preferably greater than 60 square centimeters.

The subject matter of the invention also provides a device for decanting the leukocyte content of a platelet concentrate for use in a single unit of 50 to 70 ml of platelet concentrate, wherein the effective flow area exceeds approximately 6 square centimeters.

The object according to the invention also provides a device for draining leukocyte content in a platelet concentrate including a porous, fibrous medium 12 in which the medium is an element and the device further includes a housing 11 (eng.: housing) constructed to receive the element (the element can be in the form of a right circular disc) and wherein the outer dimensions of the member are greater in lateral dimension than the inner lateral dimensions of the associated housing.

The object of the invention also provides a device for draining the leukocyte content of a platelet concentrate which includes a porous, fibrous medium 12 in which the porous medium has been preformed to form a preformed element of controlled pore diameter assembly in a housing, and in which the preformed element has been shaped or sized by compression in a softened state.

The object of the invention also provides a device for tapping the leukocyte content of a platelet concentrate including a porous, fibrous medium 12 with a CWST of at least about 90 dyne/cm in which the range between the upper and lower values used to define the CWST is about 5 or less dyn/cm. The FSA of the fibrous medium can be at least 2.5 M2 , preferably greater than 3.0 M<2> , and more preferably is greater than 3.8 M2 , e.g. in the range from 2.5 to 4.0 M<2>, preferably in the range of from 3.3 to 4.0 M<2>.

The object according to the invention also provides a device for draining the leukocyte content in a platelet concentrate including a porous medium 12 with a CWST of ,. at least of approximately 90 dyne/cm and pore diameters in the range of from 3.8 to 6 µm.

The object of the invention also provides a device for draining the leukocyte content of platelet concentrate including a modified, porous, fibrous medium 12 with a CWST of at least about 95 dyne/cm, a pore diameter in the range of from 3.8 to 6 µm, and a bulk density of less than 0.36 grams/cm^, wherein the fibers of this medium include polybutylene terephthalate and having diameters of less than 30 µm, and the medium has an effective flow area exceeding 40 square centimeters and modification of the medium has been provided using a mixture of methacrylic acid and hydroxyethyl methacrylate wherein the acid/acrylate monomer weight ratio is between 0.05:1 to 0.35:1.

Significant and novel features according to this invention which help to provide high efficiency and capacity for leukocyte removal while minimizing the loss of platelets include: (a) Whereas the previously described devices have used relatively small cross-sectional areas perpendicular to the flow path, and which are correspondingly larger with regard to depth and thus the length of the flow paths, the devices according to this invention are larger in the cross-sectional area perpendicular to the flow path and smaller in depth, with a correspondingly shorter flow path. This improvement in construction contributes significantly to that

the filter lasts longer and prevents clogging.

(b) To make the large cross-sectional area both economical and practical to construct, the porous components of the preferred devices of this invention are preformed prior to assembly to closely controlled dimensions and pore diameters to form a fully self-preserving element .

Preforming eliminates the pressure on the inlet and outlet sides of the container associated with a packed fiber system, thus making devices with a larger cross-sectional area practical. The larger cross-sectional area leukocyte collection devices made possible by preforming have longer service life, as well as at least equal or usually better leukocyte removal efficiency, equal or better platelet recovery, and less fluid retention compared to devices using fibers or fibrous meshes wrapped in a housing during installation.

(c) The preferred housing, in which the element assembly is sealed, is uniquely designed to achieve appropriate application, rapid initiation and effective air clearance, resulting in reduction of PC

which is otherwise lost within the drip container.

(d) The lateral dimensions of the members are greater than

the corresponding internal dimensions of the casing in which they are mounted. For example, if the elements are disc-shaped, the outer diameter of the element is made about 0.1 to 1$ larger than the diameter of the inside of the housing. This provides a very effective seal by means of an interference fit without loss of effective area of the elements and further contributes to the minimization of the blood back-holding volume of the assembly.

(e) While treating the filter surfaces to raise their CWST to 73 dyne/cm to 90 dyne/cm is useful in achieving rapid wetting of the filters and efficient passage of the platelet suspension, a filter treated to modify its surface to a CWST in the range of 73 dyne/cm up to 90 dyne/cm adsorb a high proportion of the platelets

which is sent through this. By modifying the filter surfaces to CWST values of 90 dyne/cm or higher, a better platelet yield is achieved. A . important feature provided according to this invention is the use of filter media with CWST values higher than 90 dyne/cm, and preferably higher than 95 dyne/cm. When using polyester-fibrous media with fiber surfaces modified to CWST of 90 dyne per cm or higher, an increase in the recovery rate, together with high efficiency in leukocyte removal, is provided. (f) The high CWST values referred to in the preceding paragraph have been provided by chemical attachment of a high number of hydroxyl groups per unit surface area of the fiber surfaces. Surfaces that have been modified in this way are expected to have a relatively low negative zeta potential when immersed in water in which the pH is within the normal range for a platelet suspension, ie 7 to 7.2. Such surfaces effectively absorb leukocytes and pass through a large amount of platelets, but are not optimal in terms of allowing free passage for

the platelets.

(g) Compared to the use of only hydroxyl modification, a considerable improvement in performance is achieved by combining with the hydroxyl groups a quantity of carboxyl groups, both of which are chemically linked to the fiber surfaces. The observed improvement in performance may reflect the development of an increased negative zeta potential on the fiber surfaces at pHs of 7 to 7.2. The product modified in this way has great ability for leukocyte removal, as well as high platelet recovery. (h) Number of carboxyl groups per unit surface area appears to have an important effect on the adhesion of the platelets to the fiber surfaces. This effect is reflected in the number of platelets recovered in the filter effluent as a fraction of the number present in the platelets before filtration. As shown in Figures 5 and 8, platelet recovery is highest at the optimal amount of methacrylic acid (MAA). As demonstrated herein, the number of carboxyl groups per unit fiber surface over the area of interest in this invention, nearly proportional to the amount of MAA in the monomeric grafting solution.

It is obvious to those skilled in the art that since adequate control to provide the product at or near the top in Figures 5 and 8 can be provided by accurate control of all the relevant factors including monomer purity, type and amount of inhibitor in the monomers, monomer age, accurate measurement of the components, the use of very pure water as a diluent, the oxygen content of the grafting solution, the amount of monomer solution to which each element of the fibrous web is exposed, and the conditions of exposure to radiation such as the radiation level and the radiation time, it is highly desirable that methods be available for the use of a test to confirm that the product really has the optimum surface content of carboxyl groups. This is particularly the case when it is necessary to change the mode of operation, for example, caused by the use of a different type or size of apparatus, and is also very useful in maintaining quality control during ongoing production. With respect to the latter, as long as this test can be performed by a relatively untrained person in minutes instead of hours or days, its value is great.

Measuring the recovery of the platelets by removing the leukocytes can, of course, be used as a test. However, this is a time-consuming and expensive process, due to the fact that for this test to provide reliable data, an average of at least about 5 and preferably 10 tests must be carried out, in which each test requires a full-time employee to be carried out, and in which each test consumes large amounts of expensive platelets.

The discovery of the dye adsorption assay (DAA) test, which can be performed rapidly with the provision of reliable results, is an important aspect of this invention.

Because the term "leukocyte removal efficiency" will be used very often below, the term has been abbreviated to "removal efficiency" or to "efficiency" and these three terms will be used interchangeably. Similarly, "platelet extraction" and "extraction" will be used interchangeably below. Figure 1 is the cross-sectional view of an example of the bottling device which is included in the present invention. Figure 2 is a top view of the inside surface of the inlet section of the tapping device shown in Figure 1. Figure 3 is a top view of the inside surface of the outlet section of the tap device shown in Figure 1. Figure 4 is a cross-sectional view of the outlet section shown in Figure 3. Figure 5 is a graphical presentation of the relationship between the percentage of platelets that are recovered after passing through a filter of preferred form according to this invention, and the relationship between the methacrylic acid monomer (MAA) and the hydroxyethyl methacrylate monomer (HEMA) in the inoculation solution used to modify the surface of the polybutylene terephthalate fibers used to produce the filter medium. Figure 6 is a graphical presentation of the relationship between platelet recovery and percentage of HEMA in the inoculation solution that is used, when the MAA:HEMA weight ratio is 0.19:1. Figure 7 is a graphical presentation of the relationship between the volume of the aqueous solution of the dye, Safranine 0, adsorbed from the solution in water, expressed as cubic centimeters of 0.012$ solution per gram 2.5 um average diameter fiber, when passing in 0.38 cm per minute through a 0.7 cm high column of fibrous medium in which the fibers have been modified by radiation grafting in a base solution containing 0.43$ HEMA together with varying amounts of MAA. Figure 8 is a graphical presentation of the relationship between the amount of MAA in the seeding solution relative to 0.43 wt-$ HEMA, and the average recovery of the platelets when 6, 8 and 10 equivalent units of the platelets were passed through a filter according to this invention. Figure 9 is a graphical presentation of the relationship between recovery of platelets when the platelet concentrate is passed through a filter according to this invention that includes fibers that have been radiation grafted in a solution containing 0.43 wt% HEMA together with varying amounts of MAA, and the volume to 0.012% The Safranine 0 solution sent before the color breakthrough expressed as cubic centimeters of the solution per gram 2.5 um diameter fiber, at 0.38 cm per minute passage through a 0.7 cm high column of fibers similar to those used in a platelet extraction test.

Material for use in the construction of the devices for removing leukocytes from platelet concentrate: Various starting materials other than fibers can be considered; for example, porous media can be cast from a resin solution to produce porous membranes or sintered powder media can be used. But taking into account price, suitability, flexibility, and ease of manufacture and control shows that fibers are a preferred starting material.

As discussed above, the blood component devices should be made from materials having CWST values in the range of about 73 dyne/cm or higher to achieve good initiation with fully wetted fibrous media. Practical considerations lead to a preference for the use of commercially available resins, synthetic resins from which fibers have been prepared commercially from include polyvinylidene fluoride, polyethylene, polypropylene, cellulose acetate, polyamides such as Nylon 6 and 66, polyesters, acrylics, polyacrylonitriles, and poly- aramids. An important characteristic of resins is their critical surface tension (Zisman, "Contact angles, wett-ability and adhesion", Adv. Chem. Ser. 43, 1-51, 1964 ). The above resins have critical surface tensions ("Yc") ranging from 27 to 45 dynes/cm. Experiments have shown that the CWST of the filters in the pore sizes required for the products of this invention is expected to be less than about 10 dynes/cm higher than "Yc. For example, for 0.2 µm pore diameter polytetrafluoroethylene -yc is 18 and CWST is 27 to 28, while for a PBT fibrous food, yc is 45 and CWST is 52. We did not find any commercially available synthetic fibers such as; after the formation of a mesh with a pore size less than 20 µm, had a CWST greater than about 52 dyne/cm.

Natural fibers less than about 15 µm in diameter are not generally commercially available. Synthetic meshes having fiber diameters less than 3 µm made by the meltblowing process have been available for about 15 years, and compared to natural fibers, such fibers require one-fifth or less of the mass to provide equal fiber surface area for leukocyte adsorption, and they additionally occupy less volume when they are used to produce the filters with a given pore size. Because of this, natural fibers are not well suited for the production of leukocyte removal devices with an optimally low retention volume. For example, a commercially available packaged cotton fiber device that has been tested for leukocyte trapping has a retention volume in excess of 75 ml, which is more than twice the volume of the preferred finished device described in this application. In addition, this device requires saline to be passed before and after the platelet concentrate is passed, and is not suitable for bedside use, and platelet extraction is only about $50. In addition, blood that has been processed in this way must be used within 24 hours.

Surface grafting has been the subject of extensive research for 25 years or more, multiple publications in the scientific literature and a large number of patents describe various methods and procedures for achieving surface modification in this way. In such a process using different monomers each containing an ethylenic or acrylic component together with another group which can be selected varying from hydrophilic (eg -COOH, or -OH) to hydrophobic (eg a methyl group or saturated chains such as —CH2CH2CH3) have been used in the development of this invention. Heat, UV and other reaction energy-giving methods can be used to initiate and complete the reaction. The ionizing radiation (7-beam) from <60>Co has been chosen to be most appropriate and has been used in this invention to modify the CWST into fibrous mats. By cutting and drying selection, mixtures of monomers or single monomers can be found which will produce a fibrous mat consisting of polybutylene terephthalate in which the CWST has been increased from 52 to any desired value up to as high as possible measurable by the method described above. The upper limit is set by the lack of liquids with surface tensions at room temperature higher than about 115 dyne/cm.

During the development of this invention, devices were made using media in which the grafting was provided by means of compounds containing an ethylenically unsaturated group such as an acrylic component combined with a hydroxyl group, for example, hydroxyethyl methacrylate (HEMA). Use of HEMA as the sole monomer produces very high CWST fibrous PBT media. Incorporation of methacrylic acid (MAA) into the monomer solution causes the zeta potential of the grafted porous medium to become more negative.

A feature which has been provided according to this invention is the use of the monomers described above, or of their analogues with similar characteristics, to modify the surface characteristics of polyester or other organic fibers to influence their behavior when contacted with the platelets and the leukocytes suspended in blood plasma.

It has been observed that some grafting monomers or combinations of monomers, when used to treat the fibrous porous structures described herein, react differently than others with respect to the range between the upper and lower values used to define CV/ST. This range can vary from less than 3 or up to 20 or more dyne/cm. In considering media for use in the test program conducted in connection with this invention, we have preferred to use media that have a range between the upper and lower values of approximately 5 or less dyne/cm. This choice reflects the greater precision with which CWST can be controlled when narrower scopes are chosen, despite the fact that media with larger scopes can also be used. Application of a narrower scope is preferred to improve productivity and quality control.

Liquids with surface tensions less than the CWST of the porous medium will spontaneously wet the medium on contact, and if the medium has continuous pores, they will flow easily through these. Liquids with surface tensions higher than the CWST of the porous medium will not flow at all at low differential pressures, but will flow if the pressure is sufficiently increased. If the numerical value of the surface tension of the liquid is only slightly above the numerical value of CWST, the required pressure will be small. If, on the other hand, the differential is large, the pressure required to induce flow will be high.

It has been discovered that when a liquid is forced under pressure and passed through a fibrous mat having a CWST lower than the surface tension of the liquid, the flow occurs in a non-uniform manner, so that some areas of the mat become dry. This is very undesirable in a leukocyte tapping device, firstly, because when only part of the porous medium allows flow the pressure drop will be higher, causing earlier clogging; secondly, due to the fact that the entire flow is sent only through a part of the available area, which in turn increases the probability of clogging; and thirdly, because only a portion of the fiber surface area provided for adsorption of or retention by filtration of leukocytes is used for this purpose and, as a result, leukocyte removal is less efficient.

Solutions to the problems involved in poor wetting of synthetic fibers: The fiber surface characteristics can be modified by a number of methods, for example, by chemical reaction involving wet or dry oxidation, by coating the surface or depositing a polymer thereon, and by grafting reactions such as is activated by exposure to an energy source such as heat, a Van der Graff generator, ultra-violet light, or to various forms of radiation, where 7-radiation is particularly useful.

As an example of these different methods, synthetic organic fibers can be coated with polymers containing at or near one end a reactive (e.g. epoxy) moiety and at the other end a hydrophilic group.

The above methods and other methods known to those skilled in the art of surface modification may be used, but radiation grafting, when carried out under appropriate conditions, has the advantage that considerable flexibility is available in the types of surfaces that can be modified, the many different reactants which are available for modification, and the systems available for activation of the required reaction. According to the object of the invention, it has been focused on 7-radiation grafting due to the possibility of producing synthetic organic fibrous media with CWST within the range of interest according to this invention, which is from 73 to 85 dyne/cm, preferably 90 dyne/cm , and higher, up to 115 dyn/cm or higher. The products are very stable, have extremely low aqueous extractable levels and, in addition, have improved adhesion between the fibers when used in preformed elements.

Selection of fiber diameter for use in leukocyte collection devices: Adsorption of leukocytes on fiber surfaces is generally accepted as the mechanism for leukocyte removal. Because the surface area of a given fiber weight is inversely proportional to the diameter of the fibers, it is obvious to one skilled in the art that veneer fibers will have a higher capacity and that the amount measured by weight of the fibers required to achieve a desired efficiency will be less if the fibers which is used has a smaller diameter.

Because of this, it has been common to use veneer fibers for leukocyte collection. Historically, as the technology required to produce smaller diameter fibrous meshes has advanced, they have subsequently been packaged in housings and used for leukocyte collection, as well as for many other purposes, such as industrial filtration. The performance of all these previously manufactured devices is inferior to the products according to this invention.

Choice of fiber for leukocyte emptying devices:

Many commonly used fibers, including polyesters, polyamides and acrylics, can be used for radiation grafting, because they have sufficient resistance to degradation by 7-radiation at the levels necessary for grafting and are of such a structure that available monomers can react with.

As mentioned above, the fiber diameters should be as small as possible. Synthetic fibers made by conventional spin-wart extrusion and drawing are not currently available in sizes smaller than about 6 µm in diameter.

Melt blowing, in which molten polymer is attenuated into fibers by a high-velocity gas stream and collected as a non-woven web and which was first described in the 1950s and reported to produce fibers with diameters as small as one micrometer, came in production in the 1960s and 1970s and has been gradually expanded over the years with regard to the lower limit of the fiber diameter with which coherent nets could be produced. In recent years, nets with fiber diameters of about 1.5 to about 2 µm have been obtained. Even smaller diameter fibers can be produced, but it is difficult to achieve these as a continuous web.

Some resins are better suited for meltblowing fine fibers than others. Resins that are well suited include polyethylene, polypropylene, polymethylpentene, Nylon 6, polyester PET»

(polyethylene terephthalate), and PBT (polybutylene terephthalate). Of the resins listed above, PBT is a preferred material due to the fact that it can also be subjected to radiation grafting and subsequent transformation into preformed elements with controlled pore size by hot compression.

PBT has been the main resin used for the development of the products according to this invention and is also the resin used in the examples. It should be noted, however, that other resins exist which can be fiberized and provided as mats or webs with fibers as small as 1.5 pm or less, and that such products, in which the CWST is adjusted as necessary to the optimum range , may be well suited for the manufacture of equally effective but even smaller leukocyte tapping devices. Similarly, glass fibers that have been properly treated can be used to make effective devices. Devices that are produced using these or other fibers in the same way and for the reasons described in accordance with this invention shall be included within the scope of this invention.

Description of an example of a bottling device:

As shown in Figures 1-4, an example of dispensing device 10 generally includes a housing 11 and a filter-adsorbent element 12. The housing 11 has an inlet 13 and an outlet 14 and defines a fluid flow path between the inlet 13 and the outlet 14. The filter-adsorbent element 12 is contained within the housing 11 across the liquid flow path and separates unwanted substances, such as red cells, gels, fat globules, aggregates and leukocytes from a fluid, such as a suspension of platelets in blood plasma, which flows through the housing 11.

The housings can be designed to receive different forms of filter-adsorbent arrays. Such a shape is, for example, a square. These and other possible forms are in principle all functional, provided that sufficient flow area is provided.

A square filter-adsorbent array would in theory provide a more economical use of the material, but would be less reliable if an interference fit seal is used in a manner described below for houses equipped with disc-shaped filter-adsorbent arrays. If the seal is provided by edge compression around the periphery, a considerable effective area is lost to the seal. Because of this, cylindrical housings with disc-shaped filter-adsorbent arrays fitted with an interference fit seal are preferred, although other shapes may be used. The seal provided by edge compression can be supplemented with a compression seal, where the compression seal can be very narrow and provide minimal compression, with very little loss of effective area.

The housings may be fabricated from any suitable impermeable material, including an impermeable thermoplastic material. For example, the housing may preferably be fabricated from a transparent polymer, such as an acrylic, polystyrene, or polycarbonate resin, by injection molding. Such a housing is not only easy and economical to manufacture, but if it is also transparent or translucent it is also possible to observe the passage of the liquid through the housing. The housings are designed to withstand normal handling during service, as well as internal pressures up to approximately 0.2 kg/cm<2>. This allows for a lightweight construction, which is a desirable feature of this invention and made possible by the use of preformed filter-adsorbent element arrays. The force required to compress the fibers of a filter-adsorbent array constructed efficiently, in which the element is not preformed but instead fabricated by packing the fibers into a housing, can be as high as approximately 70 kg for a 62 cm<2> disk, or about 1.1 kg/cm<2>, which requires heavier, larger and more expensive - house construction.

The housing can be formed in a number of configurations, but the housing 11 of the separation device 10 is preferably formed in two sections, i.e. an inlet section 15 and an outlet section 16. The inlet section 15 includes a round inlet plate 20, and the inner surface of the round inlet plate 20 defines a wall 21 which is opposite, but not in contact with. the upstream surface of the filter-adsorbent element 12.

Inlet 13 leads the fluid to an inlet plenum 22 between the wall 21 and the upstream surface of the filter-adsorbent element 12. According to one aspect of the invention, the inlet 13 leads the fluid to the inlet plenum 22 at or near the bottom of the housing 11, as shown in Figures 1 and 2 .

The inlet can be shaped in different ways. The inlet 13 of the exemplified separation device 10 includes a longitudinal inlet elevation 23. The inlet elevation 23 runs along the outer surface of the round inlet plate 22 parallel to a diametrical axis A of the housing 11, which, in use, is positioned with the diametrical axis A oriented generally vertically. The upper end of the inlet elevation 23 may be shaped as a sleeve for receiving a hollow tip 24 which is used to perforate the bottom of a bag containing the fluid, e.g. a platelet concentrate bag. Inlet 13 further includes an inlet passageway 25 which opens at the upper end of the hollow tip 24, and passes through the hollow tip 24 and inlet riser 23, and communicates with the inlet plenum 22 at the bottom of the inlet section 15.

The wall 21 of the circular inlet plate 20 includes a plurality of generally concentric circular elevations 26 defining the concentric circular depressions 27. The elevations 26 abut the upstream surface of the filter adsorber stack 12. As shown in Figure 2, the elevations 26 terminate in the lower portion of inlet section 15, which defines a passageway or access 30. The access 30 exists between the inlet passageway 25 and each circular recess 27, which allows fluid to flow from the inlet passageway 25 to the circular recesses 27. Collectively, the circular recesses 27 and the access 30 define the inlet plenum 22, which distributes the fluid delivered from the inlet passageway 25 over the entire upstream surface of the filter-adsorbent element 12. To prevent aggregates and other large bodies from blocking the flow at or near the boundary between the inlet passageway 25 and the inlet plenum 22 and, at the same time, reduce the hold-up volume in the housing 11, the depth of the inlet plenum 22 is greatest at the bottom of the housing 11 and decreases along the vertical axis A to a minimal value at the horizontal centerline of the housing 11.

The outlet section 16 of the housing 11 includes a circular outlet plate 31 and a cylindrical collar 32 extending from the periphery of the circular outlet plate 31 and to the periphery of the circular inlet plate 20. The cylindrical collar 32 is preferably integrally formed with the circular outlet plate 31 and connected to the circular feed plate 20 in a suitable manner, e.g. by adhesion or by sonic welding.

The surface of. Inside] of the circular output plate 3.1. defines a wall 33 which faces but is not in contact with the downstream surface of the filter-adsorbent element 12. The wall 33 includes a plurality of generally concentric circular elevations 34 which define concentric circular depressions 35. The elevations abut the surface downstream of the filter-adsorbent element 12. The circular recesses 35 collectively define an outlet plenum 36 which collects the fluid passing through the filter-adsorbent element 12. The depth of the outlet plenum 36 is made as small as possible to reduce the residence volume within the housing 11 without inhibiting the fluid flow.

According to another aspect, the wall 33 further includes a passageway such as a slot 40 which communicates with the outlet 14 at or near the top of the outlet section 16. The slot 40, which collects fluid from each of the circular recesses 35 and conducts the fluid to the outlet 14, preferably extends from the lower part to the top of the discharge section 16 parallel to the vertical axis A. In one configuration of separation device 10 shown by way of example, the width of the gap 40 remains constant, but the depth of the gap 40, which is greater than the depth of the outlet plenum 36, increases from the lower part to the top of the outlet section 16 along the vertical axis A. The length of the slot 40 may be equal to or less than the diameter of the housing, and the width may vary while the depth may be kept constant, or both width and depth vary.

Version 14 can be constructed in different ways. However, outlet 14 of tapping device 10 includes a longitudinal outlet elevation 41 extending along the outer surface of outlet plate 31 parallel to the vertical axis A. The lower end of outlet elevation 41 may be constructed as a pipe fitting or as a sleeve for receiving a pipe fitting or other equipment. Outlet 14 further includes an outlet passageway 42 which communicates with slot 40 at or near the top of housing 11, and which extends through outlet elevation 41, and opens at the lower end of outlet elevation.

As the platelet concentrate begins to flow through the apparatus, by filling the concentrate from the bottom and emptying at the top, the air is displaced and flows towards and out of the discharge passageway 42. By accurately constructing the apparatus shown by way of example, it has been possible to reduce and almost completely eliminate the situation in which some of the fluid reaches the area 43 adjacent to the outlet passageway 42 before all air has been removed from the inner portions of the housing assembly. In the absence of gap 40, this late air flow would entrain some of the platelet suspension into the outlet tube 42. Slot 40 allows the platelet suspension thus carried to flow into the slot, where the air is harmlessly separated from the liquid suspension. The air is then raised harmlessly to outlet 14 before the upward fluid level in slot 40 and is completely or almost completely expelled. before the liquid level reaches the top of the discharge plenum 36 and discharge passageway 42. The air becomes d thus very effectively removed from housing 11 to tapping device 10 according to the invention. For example, in a dispensing device having an internal diameter of 89 mm, an initial air volume of 20 cm<3>, and a 5 cm long slot with a width of 0.3 cm and a depth of 0.2 cm at the bottom, and 0.3 cm depth at the top, the remaining volume of air passing through the outlet after 1 to 2 cm<3> of the liquid has passed through the outlet is calculated to be less than 0.1 cm<3.>

To understand the importance of the slot and flow passage configuration, the equivalent operation of a conventional leukocyte tapping device will be described.

In conventional units, the fluid enters the top of the housing and exits the bottom. The housing of such a device cannot be directly connected by means of a nail in one with the housing, and becomes . instead usually attached by means of a tip integral with the housing, and instead is attached by means of a plastic pipe system to a tip which is used to penetrate a bag upstream from the conventional housing. A see-through drip chamber is attached by a pipe system downstream from both the conventional and the new housing, and thus to the patient. During the initiation, the conventional housing together with the drip container is turned over and the liquid is forced through the conventional housing and into the drip container. This includes the disadvantage that some pressure is lost, but, even more seriously, the fluid reaches the exit of the conventional housing and enters the drip tray while as much as 1 to 2 cm<3> or more of air still remains in the conventional the house. When 3 to 4 cm<3> of the fluid has been collected in the drip container, it and the housing are returned to the normal position, which leaves a reservoir of 3 to 4 cm<3> of fluid in the bottom of the drip container. According to the new housing of this invention, only the drip container is inverted during initiation, and only about 1 cm<3> of the fluid is collected in the drip container before it is again returned to the normal position.

The transparent drip container enables an observation of the droplet velocity through the airspace, thereby providing guidance for flow regulation. It also performs another task in that delayed air entering from the conventional housing is prevented from reaching the patient. This is the case because the delayed air displaces an equivalent volume of fluid in the reservoir or drip tray. However, the reservoir must be large enough to ensure that the delayed air never completely displaces the fluid, otherwise the air can enter the vein of the patient.

Systems that allow a considerable volume of air, e.g. 2 to 3 cm<3> reaches the drip tray after it has been returned to its normal position, making this non-reproducible. Thus, the greater the volume of the delayed air, the greater the volume of fluid that must be collected during initiation in the reservoir of the drip container. At the end of administration, much of that volume remains in the drip container and is thus wasted. By maximizing the air removal and thus using a smaller reservoir in the drip container, the dispensing device according to the present invention considerably reduces the amount of PC (which is always expensive and sometimes difficult to provide) that is lost during administration.

Filter-adsorbent element 12 preferably includes a single preformed layer as described below under the heading Preparation of fibrous elements. During the development stage, housings were constructed for use in analysis which incorporated the internal configuration described above, but which were additionally variable with respect to the thickness of the filter-adsorbent element. In this way, it was possible to analyze the filter-adsorbent elements that varied in total thickness. In each case, the distance between the tips of the elevations 26, 34 of the inlet and outlet sections was adjusted to equal the nominal overall thickness of the preformed member.

To provide an interference fit of filter-adsorbent element 12 within housing 11, the filter-adsorbent elements were cut from a pre-compressed plate to a diameter up to about 1$ larger than the diameter of the inside of the cylindrical collar 32. The filter-adsorbent elements were cut in such a way that a true correct cylindrical shape at the outer edges was maintained. This, together with the slightly oversized size, provides surprisingly good edge sealing by means of an interference fit between the outer edges of the preformed element 12 and the inner periphery of housing 11, utilizing the entire area and volume of filter-adsorbent array 12, which thus reduces the residence volume.

Preparation of fibrous elements in the preferred form:

The preferred upper limit of average fiber diameter is about 6 µm, however filters with larger fiber diameters, for example up to about 10 to 15 µm which are efficient and with useful yield can be made, but such filters have increased loss of PC due to retention within the correspondingly larger filter elements of PC which are not delivered to the patient, and which are thus less desirable. The lower limit of the fiber diameter here is the lower limit of the average fiber diameter in which fibrous spins that can be easily modified to CWST values greater than about 85 dyne/cm can be produced, about 1.5 to 2 µm. If finer fibers become available in the future, these may perform similarly or better than those used to make the products of this invention.

For use in the products 1 according to this invention, melt-blowing processes are preferred, which produce webs in which the fiber diameter size distribution is as narrow as possible.

The web is treated to modify the fiber surface, either before or after the formation of the fibrous arrangement. It is preferred to modify the fiber surfaces prior to the formation of the fibrous structure due to the fact that a more cohesive, stronger product is provided after heat compression to form an entire filter element.

The method used to modify the fiber surfaces when developing the products according to this invention is 'Y-radiation grafting', on the other hand, other known methods can be used to increase the CWST of the network.

The grafting procedure is aimed at producing nets with the highest possible CWST. CWST is preferably higher than about 90 dyne/cm, and more preferred are values higher than about 93 dyne/cm.

Monomers terminating with hydroxyl or carboxyl or other neutral or electronegative end groups are preferred. Learning or using monomers with hydrophobic or electropositive groups tends to cause adhesion and thus poor recovery of the platelets.

Surface modification has been carried out on the fibrous medium in the form of webs produced by the melt-blowing process, usually in roll form, in the form of sheets produced by compressing a single or multiple layers of the web to the desired density; and in the form of slices cut from the plates. To provide fiber surface modification, these alternative forms are immersed in an aqueous solution containing about 0.05 to 1 wt% monomer of molecular composition that includes an unsaturated reactive moiety at or near one end of the molecule, for example an acrylic group, and one or more non-reactive hydroxyl groups at the other end of the molecule. Non-branched linear compounds are preferred, such as, for example, HEMA (hydroxymethyl methacrylate); other monomers that may be used include hydroxypropyl acrylate, hydroxyethyl acrylate, and hydroxypropyl methacrylate. These are commercially available as the Rocryl™ 400 series from Rohm and Haas. These four are examples, but in general any hydroxyalkyl acrylate or hydroxyalkyl methacrylate, and various related monomers may be suitable. The solution may contain another water-soluble compound which may reduce the surface tension of the grafting solution so that the fibers are better wetted, for example 2 to 10 wt.-$, and preferably 4 to 5 wt.-$, tertiary butyl alcohol (t-BuOH), or nearly similar percentages diethylene glycol monobutyl ether, or ethylene glycol monobutyl ether. An optional third component, which is preferred to be used because it provides a significant improvement in platelet yield while maintaining a high leukocyte removal efficiency, is 0.01 to 0.2% or more of a monomer. wherein the non-reactive end of the molecule contains one or more carboxyl groups, for example acrylic acid or methacrylic acid (MAA). Other suitable monomers include unsaturated mono- or di-carboxylic acids, for example itaconic acid, or anhydrides such as maleic anhydride. All that is generally required is an ethylenically unsaturated bond together with a group having an anionic character such as a carboxyl or sulfonic acid group.

Although 7-irradiation with cobalt has been used in the development of this invention, other forms of radiation or other energy source types may be used. The TJtset time at a given radiation or at a given energy level is most easily determined by trial and error. During the development of this invention, total releases in the range of 0.01 to 1.5 megarad over periods of 6 to 60 hours have been used.

After removal from the spent inoculating solution, the web, plate, or precut mold is washed in water to remove spent inoculating solution, and then dried at any convenient temperature up to 175° to 225°C. The dried product can, after cooling, be tested by applying over a representative area drops of a test fluid with a surface tension of preferably at least approximately 90 dyne/cm, and more preferably 93 dyne/cm. Spontaneous wetting by or adsorption of droplets with a surface tension greater than about 90 dyne/cm over a period of 10 minutes indicates sufficiently high CWST.

Another test can be carried out, in which the purpose is to confirm that the content of carboxyl or carboxylic acid or other acid groups is optimal for obtaining the highest or almost the highest possible yield of the platelet. In this test, a column of the dried filter media is prepared, preferably with a height of about 0.7 cm, and a solution of a dye for the anionic surfaces is passed through the column. The dye is initially completely adsorbed from the solution by means of the column, so that the liquid itself is colourless. Passage of the dye solution is stopped when the color first appears downstream of the filter medium. The volume of dye solution passed through is proportional to the acid group content of the surface of the grafted fibers.

One or more of the layers of the dried web are heated to about 175°C to 325°C and compressed, for example between hot rollers, to form a self-supporting "sheet" of the required density, thickness and the pore diameter.

After cutting to size from the plate to form a straight cylinder, a self-retaining pre-formed filter element is provided, which is then inserted into the housing described above and the two parts of the housing are sealed, providing a finished filter assembly ready for connection at the inlet end to a device for piercing a bag of platelets, and at the discharge end to a drip container and a catheter, as a sterilization and packaging form an administration set suitable for bedside infusion of leukocyte-drained platelet concentrate to a patient.

As will be demonstrated in the examples, the effective flow area, which is defined as the area through which the platelet suspension flows, is preferably greater than about 40 cm<2>, while an area greater than 50 cm<2> is preferred , and an area greater than 60 cm<2> is even more preferred.

When used with platelets three to five days old, filters with an effective flow area much less than 50 to 60 cm<2> are much more amenable to extensive clogging before passing the full volume of 6 to 10 units of platelet concentrate. If for some reason it is preferred to use an element with an effective flow area of less than about 60 cm<2>, this can be used by using a correspondingly thicker preform as the filter element, and the frequency of clogging can be minimized by providing the prefiltration layer; but by doing this the retention volume is increased and the manufacture also becomes more difficult, and thus the use of a filter with an effective flow area smaller than 50 or 60 cm<2> is less preferred. Relatively smaller flow ranges can also be used with less likelihood of clogging when using very low density filter media, for example less than 0.05 to 0.1 g/cm<3>; on the other hand, loss of PC by retention within the filter element when using this alternative can be considerably large, due to the fact that the retention of PC within the element varies inversely with respect to the density.

The surface area of the fiber incorporated into a filter array must be commensurate with the amount of platelets to be treated. For a normal platelet concentrate containing 10^ platelets per cm<3>, the fiber surface area required is approximately 0.007 M<2>/cm<3>. For example, if the average volume per platelet unit is 56 ml, and an admixture of 8 units is to be processed, the required fiber surface area is:

With this range, the leukocyte content would be reduced on average by 99.0 to 99.9$, and the platelet yield in the filtrate would be 85 to 95$. If more than approximately 0.009 M<2 >fiber surface per cm<3> is used, the efficiency will increase towards $100, while the yield will fall. If less is used, the efficiency is reduced while the yield is increased. Fiber surface areas in the range from 0.005 to 0.02 M<3> per cm<3 >platelets is helpful. The lower part of this range is preferred if the main purpose is to achieve a high platelet yield (eg, 95 to 99$) while low leukocyte removal efficiency is accepted (eg, 90 to 99$). The higher end of the range is useful if the intent is to achieve very high leukocyte removal efficiency (eg, above 99.9$) while low (eg, as low as 50$ or less) platelet yield is acceptable.

Regular U.S. hospital practice is to use 6 to 10 units of platelets for transfusion of adults while newborns receive as little as half a unit. While the adult range can be satisfactorily filled using a filter designed for use with 8 units, pediatric transfusions must use filters with proportionally smaller fiber surface areas.

Characterization of porous media by physical characteristics: Various formulas have been proposed for use in calculating the pore diameter of a fibrous filter from the fiber diameter, the density of fibers, and the bulk (apparent) density of the filter medium. None of these have proven to be useful, because they do not take the pore size distribution into account, and the effect of changing the thickness of the filter medium is not adequately accounted for by such a formula. The most important thing is that the particle retention possibilities are not calculated correctly with such formulas.

Such a formula involves, for example, calculating the average distance between the fibres. However, the average distance between the fibers cannot be a meaningful predictor of performance, because in any fluid flow path it is the largest pores or pores present that control performance. In a fibrous mat such as that produced by meltblowing, the fibers are placed in a random manner and the pore size distribution is quite large. Other methods for the production of fibrous mats, e.g. air laying, or forming on a Fourdrinier screen, also provides wide pore size distributions. If two filters both have the same average fiber separation, thus the same average pore diameter, but one is relatively more extensive with regard to pore size distribution, this will send particles that are larger. Thus, the average distance between the fibers is a poor predictor of hard particle filtration performance, and the behavior of deformable "particles", such as leukocytes and platelets, is even more difficult to predict. In particular, this formula cannot be predictive of the products according to this invention, which are produced by the melt-blowing method. In this method, the fibers emanate from spinning warts in which molten resin is attenuated into fibrous form by means of an air current, as they impinge on and are adhered to a moving substrate, (which is then thrown). The fibers are not randomly distributed, but they are lined up parallel to the direction in which the substrate that is then thrown moves. A formula that does not calculate the degree to which the fibers are parallel cannot give meaningful results. The so-called web is then compressed by hot forming. During compression, the density is increased, and the average fiber-fiber distance is reduced perpendicular to the plane of the sheet, but remains unchanged in the direction parallel to the plane of the sheet. Various other formulas have been proposed to allow calculation of the pore diameters from data in terms of fiber diameter, fiber density and bulk density, but in over 40 years of construction in the manufacture and use of filter media, the oldest of the inventors of this invention has never found a formula useful for calculating a priori the effective pore diameter of the filters for the removal of solids from liquid suspension.

Mathematical modeling of a system to remove leukocytes while passing the platelets freely is even less likely to be successful, due to the fact that both leukocytes and platelets are physiologically active. Upon contact with various surfaces, many or all of these cells can release various enzymes, growth factors, and other active agents, and they can further change shape and become motile, in ways and for reasons that are currently only partially understood.

Furthermore, leukocyte removal from PC was achieved by adsorption, instead of filtration. As demonstrated below, leukocyte removal is less dependent on pore diameter, if at all, and the applicability of measuring the pore diameter mainly involves only the determination of the minimal diameter range by which recovery is not adversely affected, as indicated in the examples below. Under these circumstances, it was necessary in the development of this invention to use a cut and try method. These were based partly on knowledge and partly on intuition, but the development that resulted in this invention was above all empirical.

Measurement of the fiber surface area, for example by gas adsorption - popularly referred to as "BET" measurement - is a useful technique, because the surface area is a direct measure of the degree of fiber surface available to remove leukocytes by adsorption. In addition, the surface area of melt-blown PBT webs can be used to calculate the average fiber diameter as follows:

Total volume of fiber in 1 gram = cm<3>

1.38

(where 1.38 = fiber density of PBT, g/cm<3>)

nd<2>L 1

due to = (1)

4 1.38

The fiber surface area is ndL=Af (2)

d 1

Dividing (1) by (2), =

4 1.38Af 4 2 9

and d or (0.345Af)_<1>

l,38 Af Af

where L = total length of fiber per gram,

d = average fiber diameter in centimeters,

and Af = fiber surface area in cm<2>/g.

If the d units are micrometers, the Af units become M<2>/g, which will be used below.

Average fiber diameter is defined above with respect to Af; however, if one compares two of the preformed fibrous elements of this invention, one having a considerably narrower fiber diameter area than the other, the element with a narrower distribution will perform better as a leukocyte removal filter. For this reason, it is preferred that the elements have as narrow a distribution of fiber diameters as possible.

Another characteristic necessary to describe a porous medium sufficiently to allow it to be reproduced is its pore diameter (Dp). Pore diameters of the filter media were determined using the modified OSU F2 method and are given as the diameter of the hard particle by which 99.9% of the particles present were removed. The F2 test that was used to perform pore size measurements is a modified version of the F2 test developed in the 1970s at Oklahoma State University (OSU). In the OSU analysis, a suspension of an artificial pollutant in an appropriate test fluid is passed through the test filter by continuous sampling of the fluid upstream and downstream of the filter during the test. The samples are analyzed by automatic particle counters for their content of five or more pre-selected particle diameters and the ratio of upstream to downstream counts is automatically recorded. This ratio is known within the filter industry as the "beta ratio".

The beta ratio for each of the five or more diameters analyzed is plotted as the ordinate as a function of the particle diameter as the abscissa, usually on a graph in which the ordinate is on a logarithmic scale and the abscissa is a log<2-> scale. A smooth curve is then drawn between the points. The beta ratio for any diameter within the range being analyzed can then be read from this curve. The efficiency at its particular particle diameter is calculated from the beta ratio using the formula:

As an example, if beta = 100, efficiency = 99$.

Unless otherwise stated, the pore diameter Dp is indicated in the examples below, the particle diameter reaches beta = 1,000, and thus the F2 efficiency at the indicated pore diameters is 99.9$.

In the modified F2 test, efficiencies in the pore diameter range from 1 to 20-25 pm were determined using an aqueous suspension of AC fine test dust as the analyte pollutant, which is a naturally occurring siliceous dust from AC Spark Plug Company. Before use, a suspension of dust in water was mixed until the dispersion was stable. The analysis flow rate was 44 liters per minute per 929 cm<2 >f filter area.

The F2 assay provides an absolute pore diameter value, but requires several hours to run each assay. To reduce the time required, data provided from the F2 analysis was correlated with a quantified version of the "bubble point test" termed the "forward flow analysis", both of which are known to those skilled in the art of filter development and application, and the correlation was used to interpolate F2 -data, thus reducing the number of F2 analyzes required.

Characteristics in addition to Dp describing a porous medium include apparent or bulk density (p) in grams/cubic centimeter (g/cm<3>), fiber density (also in g/cm<3>), thickness (t) of the medium, specified in centimeters (cm), the effective flow area of the filter element (Ac) in square centimeters (cm<2>), the surface area of the fibers (Af) in M<2>/g, the fiber diameter distribution and the CWST in dyne/cm. Specification of these parameters defines a filter a filter-adsorbent element with predictable behavior when used for leukocyte collection: (a) Af, fiber surface per grams, when multiplied by the effective flow area, filter element thickness and density (Af x Ax x t x p) of the filter, is the fiber surface area (FSA) available within the filter element for removal of leukocytes by adsorption. Furthermore, relatively uniform fiber diameter distribution is preferred. (b) According to this invention, a filter is provided which will pass a mixture of ten units of PC which are 3 to 5 days old without sealing. When Ac is greater than 50 to 60 cm<2>, sealing with 10 units of 5-day-old PC will not occur frequently, and

clogging with fresher PC will be very rare,

or does not occur at all.

(c) Dp is optimally adjusted to be large enough that platelets are not removed by filtration. The data provided during the development of this invention show that when this condition has been satisfied, Dp can be further increased by a factor of two to four without any effect on leukocyte removal efficiency, but with increased PC loss caused by increased PC retention within the filter.

A fibrous filter-adsorbent element for leukocyte trapping of PC is defined by specifying the density and fiber diameter distribution of the fibers from which they are made, as well as Ac, Af, Dp, p, t, its CWST, and by maintaining the monomer selection rules described above .

The BET surface areas were measured after the fiber surfaces had been modified by grafting but before the hot compression to form a sheet of the medium from which the preformed elements of this invention were cut.

Examples

The platelet concentrate (PC) used in these examples was obtained from donated human blood anticoagulant mixed with CPDA-1, using procedures conforming to American Association of Blood Banks standards. The source of supply was the Greater N.Y. Blood Program in Melville, N.Y.

It is currently common transfusion practice in the United States to mix together 6, 8, or 10 units of PC, where one unit of PC is defined as the amount provided from a single blood donation, usually 400 to 500 mL. Four sizes of filter housings with effective flow areas (Ac) of 4.47, 17.8, 31.7 and 62.1 cm<2> were used in the examples that follow. These sizes are referred to below as sizes A, B, C and D respectively. Size D is preferred for use when transfusing adults.

Filters that are equal in all respects except Ac will have leukocyte removal capacities proportional to Ac, provided the assay flow rates used are proportional to Ac, thus the values of any of these quantities can be calculated from assays that are run using any other size. In order to make the comparison of the data reported in the examples more convenient, unless otherwise indicated, we have reported the values obtained by using the element of D size, or those obtained by using a smaller size and then calculating the results as would be provided with the D size.

When presenting the results in the examples, the term "efficiency" expressed as a percentage is used to denote 100 multiplied by the ratio

Cl - C2

Cl

where Cl is the leukocyte content per unit volume in PC, and C2 is the leukocyte content per unit volume in the effluent. The term "recovery" is used to indicate the efficiency of the platelet recovery, expressed as a percentage and is 100 times the ratio between the average concentration of platelets in the effluent and the concentration of platelets in the influent PC. Because the devices of this invention are designed to be used primarily with 6 to 10 units of platelets, the efficiency and recovery data for 6, 8, and 10 units are reported separately, and the average of the 6, 8, and 10 units is also indicated. This last number shows the average performance that can be expected in hospital operations.

The assay flow rate was controlled to 7 cm<3>/minute for a size D device, which is our calculations for the average normal hospital bedside design, and "clogging" as used herein is defined as the condition in which the flow through a size D device falls below 1 .75 cm<3>/minute at a pressure head of 102 cm water column.

Assay flow rate of 7 cm<3>/minute for the size D device, or the equivalent for the A, B, and C sizes, unless otherwise noted, was maintained during each assay by adjusting the pressure head between the bag and the location of the tube end, where leukocyte-drained PC was collected. If the pressure head reached 102 cm, it was then held at that level until the flow dropped to 1.75 cm<3>/minute, at which point the test was terminated and the filter was assumed to be clogged. If the final flow rate was greater than 1.75 cm<3>/minute or the equivalent for the A, B, or C size filters, all of the PC had been withdrawn from the bag of admixture and the filter had not been sealed.

Assays run using A-size filters were run four in tandem, using a single bag of six pooled PCs. The flow from the bag was divided into four equal parts, each of which was delivered to a filter of size A.

All leukocyte counts were performed by conventional canister counters, by well-trained technicians, and the data reported are the average of at least two counts from each of two technicians. In most examples, the dilution of the filtered effluent for counting was such that 1 count = 55 leukocytes. Towards the end of development, when most of the effluents showed zero leukocytes, the count ratio was made twenty-five times more sensitive by using a dilution ratio of 1 to 2.2. The efficacy data given to two decimal places were provided using the 1 to 2.2 dilution ratio.

Application of an automatic determination of the leukocyte content of the effluent from an efficient filter provides incorrect results, because automatic counters are designed to be operated in the range of normal leukocyte content in normal PC. Thus, the normal operating range of automatic counters is approximately 100 to 10,000 times higher than the levels achieved in the examples herein; thus, the automatic count data is not reliable at the low levels contained in the filter effluent. In other words, the leukocyte counts provided for the effluent from an efficient leukocyte tapping device are below the background-to-signal ratio of automatic counters. The counts must therefore be carried out manually.

Platelet counts for PC received from the blood bank are obtained using a Coulter counter model no.

ZM.

The elements used in the examples were in disk form. For the D-size element, diameters of 89.1 by 89.8 mm were compressed to 88.9 mm diameter when assembled. Similarly, the C-size elements were 63.7 to 64.1 mm discs compressed to 63.5 mm on assembly, and the corresponding figures for the B size are 47.8 to 48.1 and 47.6 mm, and for the A size 24.0 to 24.1 and 23.9 mm.

An element with a total thickness of t was fitted into a housing with an external configuration as described above, with a clearance of t between the sides of the two plenums, i.e. between the tips of the ridges 26 on the inlet plate 20 and the tips of the ridges 34 on the outlet plate 31, as shown in Figure 1.

The definition of a unit of PC as used herein is the amount of PC provided from a single 400 to 500 ml blood donation. The volume of one unit is suggested by AABB (American Association of Blood Banks) to be 50 to 70 ml, but smaller units, down to 40 ml are occasionally provided. We have calculated and applied 55 ml as the average volume for a PC unit. In order to put all the data on the same basis, so that comparisons of the data provided by the use of the elements with the different sizes can be easily performed, the flow volume data for the A, B and C sizes are calculated to D size equivalents for passing 6, 8 or 10 units with 55 ml volume per unit.

The leukocyte content of PC used in the examples of this invention varied from 300 to 2700 per cubic millimetres, with an average of approximately 1150 per cubic millimetres.

In the course of the research resulting in this invention, it was discovered that better results could be provided if meltblown webs with a narrower range of fiber diameter could be produced, and this was subsequently successfully achieved. The fiber diameter distribution used in the previous examples (1-93 and 163) according to this invention was, by scanning electron micrograph (SEM) inspection, to include a wider range compared to the SEMs of examples 94-162. The narrower size distribution is preferred, as indicated below.

Examples 1-121 and 163 were grafted using a monomer containing 0.43 wt-$ HEMA, 0.082 wt-$ MAA, and 4.7 wt-$ t-BuOH in water. In the remaining examples 122-162, the composition of the seeding solution was varied as indicated for each test category.

All the elements used in the examples were preformed into plates of controlled thickness and density, and straight circular discs were then cut from the plates, thus forming a test element.

Example numbers 1-24 are set forth in Table 1. As previously stated, according to U.S. practice of using platelets that are no older than five days. Because PC develops gels and aggregates even when stored under optimal conditions, the older the PC is, the more likely it is to cause clogging of filters. Thus, determining the adequacy when sending a PC unit without sealing is best provided by using a relatively old PC.

The data in Table 1 was obtained using PC of the indicated age, with elements produced using 6 µm diameter fibers with CWST values greater than 96 dyne/cm<2> and with the optimum content of acid monomers, . compressed to a density of 0.42 grams/cm<3>.

This group of examples demonstrates the ability to deliver a full 8 to 10 units in a majority of tests, even with post-expiration PCs. Examination of the performance with the 5-day-old PC as shown in Examples 9-24 shows that three out of 16 tests exhibit clogging just below 10 units; in bedside operation, an additional period of less than 30 to 60 minutes beyond the time at which the terminal flow rate as defined above (1.75 cm<3>/minute) was achieved would exceed 10 units; in practice this would normally be easier by increasing the bag height above 102 cm. Only one of the 16 tests may have caused some PC to be left unused in the bag. If, on the other hand, the tests had been based on the use of a smaller Ac than 62.1 cm<2>, for example 50 cm<2>, it is calculated that four of the sixteen tests would have resulted in only partial delivery. If Ac had been 40 cm<2>, the results would have been even worse, probably increasing the proportion of incomplete deliveries to about half of the tests. The preferred minimum effective flow area is approximately 60 cm<2>, 50 cm<2> is less preferred, and 40 cm<2> is even less preferred.

When a single approximately 55 ml unit of platelets is to be processed, an effective flow area greater than approximately 6 cm<2> is most preferred, due to the volume being sent being approximately one-tenth of a 10 unit PC admixture.

In table 2, data for examples 25-50 are presented. The media used were within a thickness range of 0.33 to 0.36 cm. The bulk densities p were in the range 0.42 to 0.46 g/cm<3>. The surface area Af 0.53 M<2>/g, and the corresponding fiber diameter 5.5 µm. The allele tests were run using two-day-old PCs. No sealing was detected. Average efficiency is close to 100$; the results for the ten-unit tests show an average 1000-fold reduction in leukocyte concentration, while the average reduction is 10,000-fold for the 6- and 8-unit tests. The averages of the yield for the 6, 8, and 10 unit tests are $81.3, $84.5, and $87, respectively, a total average of $84.3, or stated another way, a loss of $15.7. Because the average thickness is 0.332 cm, and an average apparent density is 0.425 g/cm<3>, corresponding to a void volume of the fluid retention within the porous medium is, on average, 0.692 x .332 x 62.1 = 14 cm< 3.> This represents a loss of PC caused by residence within the medium, based on an average eight unit PC transfusion of 14/440 x 100 = 3.2$, and increases the average loss for Examples 25-50 to 15.7 + 3.2 = 18.9$. Similar calculations can be performed for the examples below.

In table 3, examples 51-64 are presented. The media used all had a thickness of 0.19 cm, with a bulk density p of 0.43 g/cm<3>. The surface area Af was 0.67 M<2>/gram, and the corresponding fiber diameter 4.3 µm. All tests were run using a two day old PC unless otherwise stated, No clogging was detected. The average efficiencies are lower in Examples 25-50, reflecting the lower average FSA available for removal of leukocytes by adsorption, 3.4 M<2> compared to 4.7 M<2>

in examples 25-50.

In table 4 and table 5, examples 65-75 and 76-93 are presented. Fiber surface area/gram Af, fiber diameter and total fiber surface area (FSA) are substantially similar on average to Examples 51-64. The bulk density p varies respectively for Tables 3, 4 and 5 from 0.43 g/cm<3> to 0.39 g/cm<3> to 0.36 g/cm<3>. The average dividend varies successively from 82.3,. to 87.7 to 90.4$, indicating that the yield is improved as the density is reduced. For this reason, when using Af = 0.67 (4.3 µm average fiber diameter), while a density of 0.43 g/cm<3> provides satisfactory <r>results, a density below 0, 36 g/cm<3 >preferred. The corresponding pore diameters are:

Because the Table 3 conditions cause some reduction in platelet yield, it is preferred that the pore diameter is greater than 3.4 µm, and more preferred are pore diameters greater than 3.8 µm.

Filters with good efficiency and yield can be produced by using larger pore diameters. Such filters have the disadvantage that the filtering elements are larger, thus increasing the volume of PC that is retained within the filter element. Because of this, it is preferred that the pore diameter is no larger than necessary to provide maximum yield of PC, for example less than 10 to 15 µm. If the pore diameter were increased, to higher than .15 to 30 µm, some leukocytes may pass through the filter without contacting a fiber to which they could have been adsorbed, thereby reducing efficiency. It is therefore preferred that the pore diameter is not greater than 15 pm, and it is even more preferred that it is not greater than 10 pm, and even more preferred that it is not greater than 6 pm, i.e. a preferred range is from 3, 8 to 6 p.m.

In this series, the efficiency increases from 99.0 to 99.4$ between 0.43 and 0.39 g/cm<3> bulk density, and from 99.4 to 99.5$ between 0.39 and 0.36 g/ cm<3>. These efficiency improvements occur, further, when the pore size is increased, from 3.4 to 3.8 pm. These observations indicate that leukocyte removal is mainly, if not entirely, a function of surface area, and thus that the main or only mechanism of leukocyte depletion is adsorption.

In Table 6 which presents examples 94-110, preformed elements produced with larger fiber surface area, Af, with smaller fiber diameter, and with narrower fiber diameter distribution were used. Fiber surface area/gram Af was 1.1 M<2>/g (2.6 pm average fiber diameter) and average FSA was 3.2 M<2>. The thickness was similar to that of Examples 65-75, but the average bulk density of 0.232 g/cm<3> for Examples 94-110 is 40$ lower than the 0.39 average of 65-75, a change necessary to achieve that the same pore size range with the use of 2.6 pm in relation to 4.3 pm diameter fibres.

The 3.2 M<2> average FSA for Examples 94-110 is less than the 3.3 M<2> average for Examples 65-75, but the efficiency is better.

In Examples 94-110, even each efficiency measurement showed $100 removal, and the average yield was as high as $94.2. In relation to these data, earlier tests had run, using fiber media with a larger area with higher FSA values, lower and more variable efficiencies, as well as lower yield.

Loss caused by staying within the elements of Examples 94-110 is only 11.3 cm<3>, or 2.6$ based on 440 cm<3> PC, for a yield of 91.6$.

Table 7 presents examples 111-121. The media character traits are identical to those in table 6, except with regard to CWST, which is lower. These examples, like those in Table 6, are arranged in increasing density. Calculated combined average yield for ten elements including the five lowest density elements for each of Tables 6 and 7 is $93.6. The similar value for the ten elements including the five highest density elements from each of the tables is $93.7. Above the range from 0.19 to 0.32 g/cm<3>, the yield is thus not affected by density. Pore diameter variation from 3 to 4 pm thus does not appear to affect the yield either.

Examples 108-110 and 117-121, all with FSA higher than 3.8 M<2>/g, represent a preferred form of this invention. The remainder of Examples 25 through 116 represent only slightly less preferred forms of this invention, and Examples 1 through 24 are even less preferred. Despite this, all of these examples perform better than some of the commercially available filters.

The efficiency data of Tables 6 and 7 indicate that although fair results can be obtained with FSA as low as 2.5 to 2.8 M<2>, better results can be obtained with FSA greater than 3.0 to 3.3 M<2>. Although FSA as low as 2.5 M<2> is satisfactory, use of the area greater than 3 M<2> is preferred, use of areas greater than 3.4 M<2> is more preferred, and values in excess of 3.8 M<2> are even more preferred. A range of from 2.5 to 4.0 M<2> can be used, while 3.3 to 4.0 M<2> is preferred.

In Table 8 presenting Examples 122-141, the test samples are substantially similar to those in Examples 94-121, except for the monomer used for grafting. While all of the previous examples were grafted using 0.43 wt-$ HEMA along with 0.082 wt-$ MAA, Examples 122-132 were grafted using only 0.43 wt-$ hema. The benefit of combining .082 wt% MAA with 0.43 wt% HEMA can be seen by comparing the data in Tables 6 and 7 with that in Table 8:

While the Table 8 data are poor compared to the previous examples, it should be noted that the data in Table 8 are better than those that exist for any device currently commercially available, and further that devices for leukocyte collection that until now is in the trade not in any way suitable for bedside use.

While all the examples prior to those in Table 8 combined 0.43$ HEMA with 0.082$ MAA, so that the acid monomer weight ratio with respect to HEMA is 0.19, the examples in Table 9 combine 0.43$ HEMA with 0.164 MAA, for a weight ratio between MAA and HEMA of 0.38. As table 9 shows, a result of this change is a reduction in average dividend to $81.5.

In table 10, examples 155-158 are presented. In this group the acid/acrylate monomer weight ratio was further increased, from 0.38 to 0.64; as the table shows, the dividend is further affected in a negative way, resulting in 1 et. average of 58.1$.

These data are plotted in Figure 5, which shows that the optimum MAA content for use with 0.43$ HEMA is about 0.18, or more generally, a ratio in the range of .05:1 to .35:1, within a further range of .01:1 to .5:1.

Table 11 presents data illustrating the effect of changing the HEMA content of the grafting solution from 0.11 to 0.7 wt%, while maintaining the MAA content at a 0.19 wt% ratio of HEMA.

Example 158 is the average of four tests in which the HEMA content of the inoculation solution was 0.11%. Similarly, Example 159 is the average of four tests with a HEMA content of 0.22$. Sample 160 is the average of the 17 samples in Table 6, with HEMA content of 0.43$. Example 161 is the average of the sixteen samples with a HEMA content of 0.54$. Example 162 is the average of ten examples with HEMA content of 0.70$.

The data in Table 11 are plotted in Figure 6, where the platelet yield is highest in the range from 0.4 to 0.5 wt-$, and that yields higher than 90$ are provided in the range from 0.28 to 0.65 wt-$ . Because these are the most preferred and more preferred ranges, respectively, it should be noted that each of the five examples provides better yields than any device currently commercially available, as well as leukocyte depletion ranging from 300-fold to more than a 1000-fold reduction.

The tests in which example 159 constitutes the average were very slow to start, and those in examples 160 were even slower to start. Both test sets were significantly slower to initiate than any other example.

Thus, a preferred lower level for the HEMA content together with a MAA-HEMA weight ratio in the grafting monomer of 0.19:1 is 0.1 wt.$, and a more preferred lower limit is 0.2$. A preferred range is 0.28 to 0.65%, and a more preferred range is 0.4 to 0.5% by weight.

These preferred ranges may vary if parameters such as the MAA:HEMA weight ratio were varied from 0.19 to 1, or if the seeding conditions were changed, for example, by using higher or lower starting concentrations and changing the irradiation or other activating conditions. Such areas should be within the scope of this invention.

As CWST affects the platelet yield in the examples in table 11, it can be seen that CWST higher than 90 dyne/cm is preferred, and that a CWST of 95 dyne/cm or higher is more preferred.

As noted above, platelet yield peaks at 0.4 to 0.5 wt-$ HEMA content in Figure 6; and in an extraordinary case, the efficiency of leukocyte removal is also peaked in the range of 0.22 to 0.7$ HEMA, which makes this a preferred range in terms of efficiency.

As little as 0.1 wt% HEMA in the seeding solution is theoretically more than sufficient to provide a complete monomolecular coating on a 2.6 µm diameter fibrous mesh immersed therein and then subjected to activation by an external energy source. It can thus be inferred that the HEMA-MAA deposited on the fiber in the above examples need not be evenly distributed, and because of this, an excess of monomer is necessary to achieve complete coverage. More even coverage can, for example, be provided by changing the grafting method with regard to the type of the energy source, or with regard to the exposure time to the energy source, or with regard to the intensity of the source. Other monomers can also achieve the high CWST values required, along with optimum yield and efficiency. In these ways it may be possible to achieve the preferred degree of surface modification with less than 0.1$ HEMA and other than 0.019$ MAA or other cationic polymer. It should be noted that products which can be obtained in this way fall within the scope of this invention.

Example 163 presented in the following paragraphs illustrates the viability and continued effectiveness of PC processed using the devices of this invention after transfusion into a patient; it also shows that in vivo the activity of platelets after transfusion remains normal or nearly normal. The filters used in Example 163 were B-size arrays, which except for the diameter were the same as in Examples 25-50. The research on which Example 163 is based shall be published in or close to those listed below, with authors T.S. Kickler, W.R. Bell, P.M. Ness, H. Drew and D.B. Pallets; The Johns Hopkins University School of Medicine, Baltimore, MD and the Pall Corporation, Glen Cove, NY, U.S.A.: Removal of leukocytes (WBC) from platelets can reduce alloimmunization to WBC antigens. We evaluated a new surface-modified fibrous polyester filter that requires no special processing of pooled platelet concentrates and can be used at the bedside. Other studies were designed to measure WBC clearance, platelet function, in vitro platelet function, and in vivo platelet survival. WBC counts were performed manually and showed a mean clearance of 99.7$ ± .56, n = 38. Platelet yield was 85.4$ ± 5.4, n = 38. Seal reversal and phase microscopy morphology were not affected. When using epinephrine, ADP, collagen, and ristocetin in platelet aggregation studies on 15 samples, there was no difference in the pre-filtration samples compared to the post-filtration samples.

111-Indium platelet survival studies were performed using autologous platelet concentrates in 5 volunteer subjects. Immediately after preparation, the platelets were filtered and labeled with 111-Indium oxin. After transfusion, samples were collected for 10 days and the results were analyzed using the gamma function curve fitting model. Platelet lifetimes for the 5 donors were 8.2, 8.1, 7.0, 9.2, 8.5 days (normal: 8.7 ± 1.1 days).

These studies indicate that filters effectively remove WBCs without significantly reducing platelet counts or altering platelet function or survival. This device offers potential that includes reducing platelet transfusion reactions and alloimmunization.

The products of this invention have been shown to be easy to use due to their ability to be used at the bedside, easy wetting ability and thus rapid onset, very little loss of PC caused by residence within the device, together with extraordinarily high leukocyte clearance. efficiency and platelet yield. Additionally, it has been demonstrated in vivo that the platelets that have been passed through the device and into a patient suffer no measurable loss in effectiveness, nor loss of their normal lifespan in the human body.

In the preceding sections of this application, filters were prepared by subjecting otherwise equally fibrous media to seeding solutions containing in each case 0.43 wt-# of HEMA, but in which the MAA varied over a wide range, from zero to 0.28 wt-# . These filters were all used in a number of platelet concentrate filtration tests and the average proportion of platelets recovered was determined, along with the average efficiency of leukocyte removal. Thus, it was determined that the most beneficial result was provided with a weight ratio of MAA to HEMA in the range of approximately .05:1 to .35:1, or more generally within the wider range of .01:1 to .5 :1.

Testing the grafted product by passing the platelet concentrate and analyzing the effluent is a satisfactory method of exploration to determine the optimum range, but this method is not suitable for use in routine manufacturing quality control due to the following reasons: (a) Platelet concentrates are very expensive , multiple tests are necessary, and many devices are required for each test. For example, if each test were run using a pooled amount of 10 units of platelets, the cost of purchasing platelets at 1988 US prices would exceed $400 per unit. test. In addition, each test requires a considerable investment in terms of work, close to one person-day per test. (b) Because of the variability in the platelet concentration,

in which no two units are exactly alike, now a large number, e.g. at least five or preferably more

are run to obtain meaningful mean values for platelet recovery efficiency.

Thus, a rapid and economically performed test that is in close agreement with platelet recovery performance is desirable, and such a test is provided in accordance with this invention.

As indicated above, addition of MAA to HEMA in the seeding solution caused the product to have a higher negative zeta potential compared to material only seeded with HEMA. Thus, zeta potential measurement can be used for the calculation of grafted fibrous media. The zeta potential can be measured by flow potential, or by suspending the fiber fragments in an electrolyte and then using a microscope to measure the migration speed of the fibers in an electric field. Both methods require trained personnel, are very slow, and often produce inconsistent data.

One reason for the discrepancy in the zeta potential measurements is that when PBT fibrous medium is grafted using only HEMA as monomer, the product was negative zeta potential. The effect on zeta potential of adding MAA to the inoculation solution is thus incremental, rather than absolute. This reduces the sensitivity of these measurements.

We have discovered a simple analytical method that can be performed quickly and that correlates with the MAA content of the grafting solution and with the platelet recovery ability of the grafted fibrous PBT. The basis of this method is the passage through a column of known weight of the porous medium of a solution of a dye that reacts with compounds containing anionic groups on their surfaces. Such a dye should preferably have a high degree of absorption or reflectivity for a visually detected or photometrically measured wavelength of light. The absorbed wavelength must be in the visible region of the spectrum, if visual observation is criteria, but may be in the ultra-violet or infrared spectrum if spectrophotometry is used to detect the presence of the dye.

In both cases, a solution of the dye is passed through a column of the filter medium, at a rate sufficiently low to allow the dye solution to equilibrate with the seeded medium. If the selected dye has a high positive zeta potential, such as dyes where the active group or groups may be amine, or preferably quaternary ammonium group or groups, it will be ionically adsorbed to carboxyl or other anionic groups on the surfaces of fibers. When flow is first initiated, the dye is completely adsorbed and clear liquid, especially pure water, emerges. When all surface groups near the entrance to the column have adsorbed dye molecules and are saturated, the dye moves to the next level of the column, which is similarly saturated, and this process continues until the solution containing the dye runs out the end of the column. Occurrence of color, or absorption of light waves by the dye if spectrophotometry is used, signals that the contents of the column have been saturated. The volume of dye that has been sent at this time is a measure of the surface population of carboxyl or other anionic groups at the surface of the porous medium.

As the dye progresses through the column, the front has a considerable bandwidth, caused by the time required for diffusion to occur from the dye solution to the fiber surfaces, as well as caused by diffusion of the dye in a vertical direction. This bandwidth can be reduced to less than approximately 1 mm in width by choosing an appropriate flow rate and dye concentration, and by evacuating the column before the first filling with liquid. If there is any possibility that the medium being tested has been exposed to metal ions during the grafting process and subsequently washed free of residual grafting fluid, the media used for this test should be returned to the acid form by exposure to a weak acid solution, and then washed free of acid using deionized water. To ensure reproducible results, the column height should be at least 0.7 cm and the density of the column should be approximately 0.23 grams per cm<3>.

The column can be produced from porous medium in sheet form by cutting an appropriate number of disks and assembling these into a preferably transparent tube with an inside diameter of approximately 1 to 1.5 cm. Discs must be cut so that their outer edge is perpendicular to the plane of the disc, each disc forming a true right cylinder approximately 1$ larger than the inside diameter of the transparent tube.

A number of dyes well suited for this purpose are available. We have used a quaternized diamine, Safrinine 0, with the following structure:

Safranine 0 is used as a biological dye, and is thus available in many laboratories, and is therefore an appropriate choice.

Examples 164-167 show the ratio between the volume of a .012$ Safranine 0 solution adsorbed per grams of fiber, and proportion of MAA based on the HEMA content added to 0.43 wt-$ HEMA in the grafting solution used to modify the fiber surface. The resulting data is shown in Figure 7. This relationship appears to be linear.

Media prepared identically to those used in Examples 165 to 167 were seeded in a manner similar to the filters used in Examples 111-154, except that: (a) the average fiber surface area was 3

square meters,

(b) average fiber diameter was 2.5 µm,

(c) the average density was 0.22 g/cm<3>,

(d) the fiber size distribution was somewhat narrower,

(e) the radiation grafting was carried out in the production-scale apparatus, compared to the laboratory scale used

in the previous examples,

(f) HEMA content remained at 0.43 wt-$, while MAA:HEMA

the weight ratio was varied to include .064, .13, .1.9 and .25$.

The resulting filters were all used in twelve or more tests at each MAA-HEMA ratio, to pass platelet concentrate equivalent to 6, 8 and 10 units respectively, providing for each test three values for platelet yield, and three for leukocyte removal efficiency. The three platelet yield percentages provided in each of the twelve or more tests were averaged and the results are tabulated in Table 13. The leukocyte clearance averages were similarly determined and all were essentially the same, and the range of the four different MAA concentrations were $99.84 to $99.91.

The platelet yield data for Examples 168-171 were plotted against the proportion of MAA used in the seeding solution relative to 0.43$ HEMA (Figure 8), and shows a preferred ratio of MAA to HEMA in the seeding solution relative to the HEMA content of 0.43$ and is in the range .05 to .25.

The minor difference in preferred area between the data in Figures 5 and 8 is believed to be caused by not unexpected changes in the product that are provided when going from laboratory scale (Figure 5) to production scale (Figure 8), and may also reflect the application of fibers with a narrower size distribution. Both Figure 5 and Figure 8 show that a ratio of approximately 0.19 is optimal. This difference underlines one of the advantages of controlling the product characteristics by means of the dye adsorption method.

The platelet yield data of Examples 168-171 plotted against Safranine 0 adsorption in Figure 9 indicate a preferred range of 10 to 35 cm<3> of 0.12$ Safranine 0 pr. gram, and a more preferred area of 17 to 34 cm<3> of .012$ Safranine 0 per gram.

(<!>) Mixed PC, 50$ each of the units 7 days and 8 days old.

Claims (38)

1. Device for reducing the leukocyte content in a platelet concentrate or solution, or treating blood or blood component comprising a housing (11) with an inlet (13) and an outlet (14) and a fluid flow path between the inlet (13) and the outlet (14), characterized in that the fluid flow path is a porous, fibrous medium (12) with a CWST of at least 90 dyne/cm and a negative zeta potential. 2. Device according to claim 1, characterized in that said medium has a CWST of at least 95 dyne/cm, and a negative zeta potential. 3. Device according to claim 1, characterized in that the fibers of the medium have been modified so that they include hydroxyl groups when they are immersed in an aqueous fluid. 4. Device according to claim 3, characterized in that the fibers have been modified by contact with a monomer comprising a polymerizable group, preferably comprising an acrylic or methacrylic part and a hydroxyl-containing group, preferably comprising hydroxyethyl, under polymerization conditions. 5 . Device according to claim 4, characterized in that said monomer is hydroxyethyl methacrylate. 6. Device according to claim 1, characterized in that the fibers of the medium have been modified so that they present a hydroxyl group together with a smaller number of another anionic group, preferably a carboxylic acid group. 7. Device according to claim 6, characterized in that the fibers of the medium have been modified with a monomer containing a polymerizable group, preferably including an acrylic or methacrylic part, and a carboxyl-containing group under polymerization conditions. 8. Device according to claim 7, characterized in that the polymerizable group and the carboxyl-containing group are provided by means of methacrylic acid. 9. Device according to claim 8, characterized in that the fibers of the medium have been modified using a mixture of monomers including methacrylic acid and hydroxyethyl methacrylate. 10. Device according to claim 9, characterized in that the acid/acrylate monomer weight ratio in the modified mixture is between 0.01:1 and 0.5:1. 11. Device according to claim 10, characterized in that the acid/acrylate monomer weight ratio in the modified mixture is between 0.05:1 and 0.35:1. 12. Device according to claim 7, characterized in that the modification has been carried out using a grafting solution containing 2 to 10% by weight of tertiary butyl alcohol. 13. Device according to claim 12, characterized in that the modification has been carried out using a grafting solution containing 4 to 5% by weight of tertiary butyl alcohol. 14 . Device according to claim 7, characterized in that the modification has been carried out using a grafting solution containing sufficient water-soluble alcohol, or alcohol-ether to reduce its surface tension to less than 40 dyne/cm. 15 . Device according to claim 14, characterized in that the alcohol ether is diethylene glycol monobutyl ether or ethylene glycol monobutyl ether. 16. Device according to claim 10, characterized in that the concentration of hydroxyethyl methacrylate in the modified mixture is higher than 0.1% by weight. 17. Device according to claim 16, characterized in that the concentration of hydroxyethyl methacrylate in the modified mixture is in the range from 0.2 to 0.7 by weight. 18. Device according to claim 1, characterized in that at least one element of the porous medium has been preformed before assembly in a housing. 19. Device according to claim 18, characterized in that the preformed element includes organic fibers with less than 30 µm in diameter. 20. Device according to claim 1 or 2, characterized in that the pore diameter is greater than 3 µm. 21. Device according to claim 20, characterized in that the pore diameter is greater than 3.4 µm. 22. Device according to claim 21, characterized in that the pore diameter is greater than 3.8 µm. 23. Device according to claim 20, characterized in that the bulk density of the medium is less than 0.36 g/cm<3>. 24. Device according to claim 1, characterized in that it is to be used with a mixed amount of 6 to 10 units of platelet concentrate, each with a volume of 50 to 70 ml and in which the effective flow area is greater than 40 square centimeters. 25 . Device according to claim 24, characterized in that the effective flow area is greater than 50 cm<2>. 26. Device according to claim 24, characterized in that the effective flow area is greater than 60 cm<2>. 27. Device according to claim 1, characterized in that it is to be used with a single unit of 50 to 70 ml of platelet concentrate, in which the effective flow area is greater than 6 square centimeters. 28. Device according to claim 1, characterized in that said medium is an element, preferably shaped like a straight circular disc, and said device further includes a housing designed to receive the element and in which the outer dimensions of the element are larger in lateral dimension than the inner lateral the dimensions of the corresponding house. 29. Device according to claim 1, characterized in that the porous medium has been preformed to form a preformed element with controlled pore diameter before assembly in a housing. 30. Device according to claim 28, characterized in that the preformed element has been shaped or made in one certain size by compression in a softened state. 31. Device according to claim 1, characterized in that the range between the upper and lower values used to define CWST is 5 or fewer dyne/cm. 32. Device according to claim 1, characterized in that the FSA of the fibrous medium is at least 2.5 M<2>. 33. Device according to claim 32, characterized in that the FSA of the fibrous medium is from 2.5 to 4.9 M<2>. 34. Device according to claim 1, characterized in that the pore diameter is in the range from 3.8 to 6 µm and a negative zeta potential. 35. Device according to claim 1, characterized in that it has a CV/ST of at least 95 dyne/cm, a negative zeta potential, a pore diameter in the range of from 3.8 to 6 µm, and a bulk density of less than 0.36 grams /cm<3>, the fibers of said medium include polybutylene terephthalate and have diameters of less than 30 µm, the medium has an effective flow area greater than 40 cm<2> and the modification of the medium has been provided by using a mixture of methacrylic acid and hydroxyethyl methacrylate wherein the acid/acrylate monomer weight ratio is between .05:1 to .35:1. 36. Device according to claim 1, characterized in that the bulk density of the medium is less than 0.43 g/cm<3>. 37. Device according to claim 1, characterized in that the bulk density of the medium is in the range from 0.19 to 0.43 <g>/cm3. 38. Device according to any one of the preceding claims, characterized in that the device includes at least one prefiltration layer upstream of the fibrous medium in the same direction as the fluid flow.